/SR-GNN

A PyTorch implementation of SR-GNN [AAAI 2019] "Session-based Recommendation with Graph Neural Networks"

Primary LanguagePython

SR-GNN

This is a pytorch implementation of AAAI'19 paper Session-based Recommendation with Graph Neural Networks

Running on single GPU:

python main.py

Performance

  • Result was slightly better than original repo on sample dataset.
  • This repo also uses the exact network architecture proposed in the paper, which has a different GNN aggregation logic to the official repo.
  • Speed is slightly slower than the original repo because of log file writing.